The present invention relates to an improved method and apparatus for the continuous culture of biocatalysts. More particularly, the present invention relates to a method and apparatus for culturing micro-organisms, or plant or animal cells, or subcellular cell components as three-dimensional arrays immobilized in centrifugal force fields which are opposed by liquid flows. The present invention allows the maintenance of extremely high density cultures of biocatalysts and maximizes their productivity.

The term "fermentation" as used herein means any of a group of chemical reactions induced by living or nonliving biocatalysts. The term "culture" as used herein means the suspension or attachment of any such biocatalyst in or covered by a liquid medium for the purpose of maintaining chemical reactions. The term "biocatalysts" as used herein, includes enzymes, vitamins, enzyme aggregates, immobilized enzymes, subcellular components, prokaryotic cells, and eukaryotic cells. The term "centrifugal force" means a centripetal force resulting from angular rotation of an object when viewed from a congruently rotating frame of reference.

The culture of microbial cells (fermentation) or animal and plant cells (tissue culture) are central to a multiplicity of commercially-important chemical and biochemical production processes. Living cells are employed in these processes as a result of the fact that living cells, using generally easily obtainable starting materials, can economically synthesize commercially-valuable chemicals.

Fermentation involves the growth or maintenance of living cells in a nutrient liquid media. In a typical batch fermentation process, the desired micro-organism or eukaryotic cell is placed in a defined medium composed of water, nutrient chemicals and dissolved gases, and allowed to grow (or multiply) to a desired culture density. The liquid medium must contain all the chemicals which the cells require for their life processes and also should provide the optimal environmental conditions for their continued growth and/or replication. The living cells, so maintained, then metabolically produce the desired product(s) from precursor chemicals introduced into the nutrient mixture. The desired product(s) are either purified from the liquid medium or are extracted from the cells themselves.

There are a number of disadvantages inherent in typical fermentation processes. Such processes require expensive energy expenditures to maintain the large volumes of aqueous solution at the proper temperature, the process must be continuously monitored and additions made to maintain nutrient concentration and pH at optimal levels, and extreme care and expense must be taken to sterilize and then to exclude undesired cell types from gaining access to the culture medium. Such fermentation methods, particularly those employing aerobic organisms, are limited to low yields of product or low rates of product formation as a result of the inability to deliver adequate quantities of dissolved oxygen to the metabolizing organism. Finally, such batch or semi-batch processes can only be operated for a finite time period before the buildup of excreted wastes in the fermentation media require process shutdown followed by system cleanup, resterilization, and a re-start.

Cell immobilization allows for a much greater effective density of cell growth and results in a much reduced loss of productive cells to output product streams. An early method for the immobilization of cells or enzymes involved the entrapment of such biocatalysts on dextran, polyacrylamide, nylon, polystyrene, calcium alginate, or agar gel structures or spherical polymeric microcarrier beads. These immobilization methods suffer from a number of drawbacks, such as interference with the diffusion of gases, poor mechanical properties and high compressibility of gel-entrapment media leading to high pressure problems, the crushing of microcarrier beads and the destruction of attached cells by hydraulic shear forces in agitated tank reactors.

US-A 4,939,087 discloses a bioprocessing apparatus and method using centrifugal acceleration of cells to maintain a culture within a reaction zone inside of the reactor. A continuous flow of a liquid is maintained in the reactor, which acts to create force which opposes the centrifugal force field. The reactor is placed on a horizontal rotating plate and is rotated horizontal around an axis in the middle of the horizontal plate. Therefore the gravitational force is perpendicular to the centrifugal and the liquid flow forces in the reactor. Thus, during a long term centrifugation the gravitational force causes the particles to settle in the lowest region of the centrifugal reactor.

JP-54114474-A describes a rotary container with a perforated plate on its outside periphery, whereby the rotation center of the container is provided with an outlet pipe. The container is rotated to closely pack the particles contained inside against the outside periphery of the container. A rotation axis is contained in the middle of the centrifugal container, resulting in a centrifugal force perpendicular to the gravitational force. A fluid is introduced into the container through a perforated plate from outside, resulting in a liquid force perpendicular to the gravitational force inside of the container.

Another method for the immobilization of living cells or enzymes currently in use involves the use of packedbed reactors. In these methods, free cells or cells bound to microcarrier beads are suspended in a rigid or semi-rigid matrix which is placed within a culture bioreactor. These methods of immobilization all suffer from a number of problems, particularly when scaled up to production size.

A more recently-developed class of methods for cell immobilization involves the confinement of the desired cells between two synthetic membranes. Such processes provide the cells with an environment in which nutrient liquid input and waste liquid output can occur through channels separate from the cell-containing space and similarly provide gaseous input and output through similarly disposed channels, again separate from the cell-containing space. Unfortunately, there are a number of problems with such methods, particularly for any commercial, large-scale usage, such as the membranes are costly to manufacture, difficult to assemble, cells attach across pores and/or pore clogging by insolubles in the nutrient feed or waste output liquids which wet this membrane, and a large number of costly inlet and outlet ports and external fittings.

Another class of methods for cell immobilization involves the employment of capillary hollow fibers (usually configured in elongated bundles of many fibers) having micropores in the fiber walls. Typically, cells are cultured in a closed chamber into which the fiber bundles are placed. Nutrient aqueous solutions flow freely through the capillary lumena and the hydrostatic pressure of this flow results in an outward radial perfusion of the nutrient liquid into the extracapillary space in a gradient beginning at the entry port. Similarly, this pressure differential drives an outward flow of "spent" media from the cell chamber back into the capillary lumena by which wastes are removed. Cells grow in the extracapillary space either in free solution or by attachment to the extracapillary walls of the fibers. Typically, oxygen is dissolved into the liquid fraction of the extracapillary space by means of an external reservoir connected to this space via a pump mechanism. Waste products in the intracapillary space may be removed by reverse osmosis in fluid circulated outside of the cell chamber. There are a number of difficulties with the use of methods based on capillary hollow fiber cell immobilization methods. Cracauer et al. (U.S. Patent No. 4,804,628) have extensively documented these difficulties.

Another class of methods for the mass culture of living cells involves the use of fluidized bed bioreactors. The major disadvantage of fluidized bed methods, and particularly a variant called airlift fermentors, results from the necessity to bubble air or oxygen through the reactor and the resultant presence of a gas-liquid interface throughout the reactor volume. The presence of gas bubbles disrupts the fluid dynamics and protein foaming, cell destruction, and the denaturation of nutrients and products occurs at the large gas-liquid interface. Cell washout is almost inevitable in continuous operation, particularly with animal cell culture.

Another class of methods for mass cell culture is known as dual axis, continuous flow bioreactor processing. In this class of bioreactor, rotation of the reactor chamber about an axis perpendicular to the vertical axis is utilized in order to effect internal mixing of the bioreactor contents while rotation about the vertical axis confines grossly particulate matter at radial distances far from the vertical axis of rotation. Methods of this class are ineffective in the immobilization of low mass micro-organisms, particularly those requiring gaseous nutrients and producing waste gas products.

Another type of bioreactor called a "Nonhomogeneous Centrifugal Film Bioreactor" intended for aerobic cell culture is taught by U.S. Patent No. 5,248,613. The object of the method is to maximize the "entrainment of the maximum amount of the gaseous phase into the liquid phase" by causing the formation of a thin liquid film to contact the gas phase and further, to centrifugally generate small liquid droplets which fall through a relatively stationary gas phase back into the recirculated bulk liquid phase. A final method for the mass immobilization of living cells called "Continuous Centrifugal Bioprocessing" has been taught by Van Wie, et al., U.S. Patent No. 4,939,087. In this method cells are "captured" by a velocity gradient in a centrifugal field in order to maintain a culture in a revolving bioreactor chamber into which and out of which liquid flows are pumped. As is shown in Figure 1, the basic idea is to suspend a particle in a spinning bioreactor chamber, which as a consequence of its rotation, imparts a "centrifugal" force to the particle which would normally cause the particle to migrate to longer centrifugal radii. Liquid flow is introduced into the periphery of the spinning chamber (and withdrawn at shorter radii) in order to impart an opposing force which counteracts that of the centrifugal field, with the result that the particle is immobilized at a particular radial distance in a liquid flow. The essence of Sanderson and Bird's (Sanderson, R.J. and Bird, K.E. (1977) Methods in Cell Biology, 15, 1-14) mathematical analysis of the particle and fluid dynamics of this process are displayed in Figure 2.

This theory, when applied to Centrifugal Elutriation, as it was by Sanderson and Bird and by developers at Beckman Instruments, can be utilized in the short term for the separation of cells of different size and/or density. Unfortunately, it is completely inapplicable to long term immobilization of cells or biocatalysts since the theoretical basis is incorrect. As is shown in Figure 3, there is an additional force acting on the suspended particle which must be taken into account, particularly when the particle is to be immobilized over long time periods (as would be the case in fermentations). That force is a result of the particle's mass. Whereas micro-organisms or animal cells are quite light in weight, their mass is non-zero and consequently gravity will have an effect which is significant, the more so over longer time periods. This is shown more graphically in Figure 4, where it is shown that there is not a simple description of a radial distance where a particle in an applied centrifugal field can be immobilized in a flowing liquid since the derivation has neglected to consider the effect of gravity on the mass of the particle. The result of this "deviation from theory" is evident in centrifugal elutriation experiments which require prolonged separation times and is shown graphically in Figure 5. Over longer time periods, the weight of the suspended particles (shown as dark circles in a circular cross-section of a bioreactor chamber in Fig. 5) will cause these particles to settle to the lowest regions of the bioreactor chamber, disrupting the balance of forces which initially suspended them in the chamber. Further, the "aggregation" of these particles into a larger "particle" with virtually the same density as the individual particles results in an increased centrifugal effect which causes the aggregates to migrate to longer radii, eventually clogging the liquid input port.

The prior art demonstrates that while cell immobilization is a greatly desired method for increasing the productivity of living cells in culture, there are a number of drawbacks associated with each class of method. A central problem of all such culture methods is adequate oxygenation of the cultured cells and removal of carbon dioxide has been a limiting factor in the development of more efficient and economical designs.

Living cells or bio-catalytic subcellular components are unable to derive any benefit from gaseous oxygen. Living cells or biocatalysts derive benefit solely from oxygen dissolved within the aqueous media which surrounds the particles. In batch fermentations which are common for microbial production, the sparging of air or oxygen-enriched gases through the aqueous nutrient media is intended to replace the dissolved oxygen consumed by the metabolizing cells. In this method, most of the gas exits unused while dissolved oxygen levels are maintained at some value. Similarly, the sparging of air (or oxygen) into the nutrient media prior to its use in animal cell culture is intended to maintain a level of dissolved oxygen in the media. While the normal concentration of oxygen in water varies from about 0.2 to 0.3 mM (depending on such factors as pH and ionic strength), it is possible to increase this concentration to as much as 0.5 mM by applying approximately two atmospheres of oxygen pressure over a water solution.

To maintain adequate oxygen concentrations in fermentation media, most of the prior art has focused on increasing the contact between gas and liquid by: (1) producing a very small bubble size (a function of the sparging frit pore size); (2) using high speed agitation to increase the rate of oxygen entrance into the liquid phase; or (3) using a gaseous overpressure of one or two atmospheres above the culture medium to increase dissolved oxygen levels. In the case of animal cell culture, the typical design of animal cell culture chambers has heretofore made it difficult to consider using overpressures greater than a fraction of an atmosphere. Thus, the most common method for increasing oxygen levels employs gas permeable membranes or fibers in contact with flowing nutrient liquid to maintain dissolved oxygen levels.

There are a number of problems associated with these methods of increasing the concentration of dissolved oxygen in nutrient media. First and foremost, nearly all of these methods are unable to increase dissolved oxygen concentrations above that obtainable at atmospheric pressure due to the generally fragile nature of other components of the cell culture process. Next, methods which involve vigorous agitation of the liquid-gas mixture to effect increased rates of oxygen dissolution are not applicable to animal cells, which are quite fragile and can easily be damaged by hydraulic shear forces. Finally, those methods which do apply an increased gaseous overpressure above the culture media to increase dissolved oxygen concentrations cannot be scaled up much more than to approximately 1-2 atmospheres of overpressure before it becomes impossible to access the cell-containing liquid media for cell harvest or product isolation without destroying the cultured cells. Nevertheless, the teachings of each of the above methods warrant individual discussion.

U.S. Patent No. 4,897,359 (issued to Oakley, et al.) discloses a method for oxygenating animal cell culture media for subsequent introduction into cell culture vessels in which an oxygenated gas, at an indeterminate pressure, is passed through a multiplicity of gas permeable tubes surrounded by the liquid medium to be oxygenated. The method of the invention of Oakley, et al. is useful only in assuring that the cell culture media possesses the maximum dissolved oxygen concentration obtainable at atmospheric pressure.

Finally, because the immobilization of cells or microorganisms requires that a cell confinement chamber is part of the process system, the recent literature has been examined for comparison. There are a number of cell culture chambers in existence. Many of these chambers provide for the input and output of a liquid stream, several have viewing ports, and all provide a surface upon which cells may attach or a chamber in which suspended cells may be cultured. In all cases, the operating pressure of these confinement chambers is one atmosphere (or less) and thus these chambers are unsuitable for processes in which increased dissolved oxygen levels are desired and are necessarily limited to those dissolved oxygen levels obtainable at atmospheric pressure.

The current state of the art reveals that there are three inter-related problems which plague the economic use of mass cultures of microbes, animal cells, or their subcellular components: (1) As is evident from the sheer volume of the prior art on cell immobilization, the focus of much work has been to increase the density of the cell culture. It is obvious that the economical production of a biological product will be directly related to the ability to efficiently culture large aggregates of the desired cell type. Unfortunately, the drive to increase cell culture density leads to the evolution of the two secondary problems: (2) the inability to adequately nutrition a high density cell aggregate; and (3) the inability to supply adequate oxygen to high density aerobic cell populations. As cell density is increased, the only method for the supply of adequate liquid nutrient to the aggregate involves increased liquid flow rates which, in all cases in the prior art, eventually limits the overall scale of the immobilization method. Similarly, as the cell density increases, the inability to deliver adequate dissolved oxygen (or any other gas) to the cell aggregate is even more of a limiting factor and severely reduces the scale of the culture.

Accordingly, there remains a need for an apparatus and method for continuously culturing, feeding, and extracting biochemical products from either microbial or eukaryotic cells or their subcellular components while maintaining viable, high density aggregates of these biocatalysts. In addition, there is a need for a method for the absolute immobilization of sample biocatalyst populations which will allow the study of various nutritive, growth, and productive parameters to provide a more accurate understanding of the inter-relationships between these parameters and their effects on cell viability and productivity.

The present invention comprises a novel culture method and apparatus in which living cells or subcellular biocatalysts are immobilized within bioreactor chambers mounted in a centrifugal field while nutrient liquids without any gas phase(s) in contact with the liquids are flowed into and out of the immobilization chambers. The cells or biocatalysts are ordered into a three-dimensional array of particles, the density of which is determined by the particle size, shape, intrinsic density, and by the selection of combinations of easily controllable parameters such as liquid flow rate and angular velocity of rotation.

According to the present invention, the cells or biocatalysts can be confined within the bioreactor chambers at a defined volume. Only liquids (which may contain dissolved gases) are passed into and out of the chambers. To cause nutrient liquids to flow through the three-dimensional array of cells or catalysts in the chambers, positive displacement pumps are employed to move the nutrient liquid, at positive hydraulic pressure, through the chambers. The confined cells or biocatalysts are unaffected by the resultant increase in hydraulic pressure as long as high frequency pressure fluctuations are not present. Thus, fresh, optimal liquid nutrient media is presented to the confined cells or biocatalysts at all times during the process flow while desired cellular products are immediately accessible at the output of the confinement chambers.

The present invention can be used to produce high yields of industrial chemicals or pharmaceutical products from biocatalysts such as bacteria, yeasts, fungi, and eukaryotic cells or subcellular organelles, such as mitochondria, or immobilized enzyme complexes. These cells or cellular substructures can be either naturally occurring or can be genetically manipulated to produce the desired product. The present invention can be operated in either of two modes: (1) a mode in which nutrient limitation is used to insure a defined bioreactor bed volume. This mode is applicable to cultures where desired products are released from the immobilized biocatalysts and exit the bioreactor in the liquid flow; (2) a mode in which excess nutrient input is used to cause overgrowth of the volume limitation of the bioreactor. This mode is useful for the continual production and outflow of mature cells containing an intracellular product.

Accordingly, it is an object of the present invention to provide a method and apparatus by which biocatalysts are immobilized within centrifugal immobilization chambers while nutrient liquids are fed into the chambers and effluent liquids containing desired metabolic product(s) exit the chambers.

It is a further object of the present invention to provide a method and apparatus by which biocatalysts, including living cell populations, may be immobilized and either aerobic or anaerobic fermentations performed in which liquid nutrient and substrate nutrients are converted to product-containing output liquid streams.

It is a further object of the present invention to provide a method and apparatus by which bacterial cell populations may be immobilized and fermentations performed in which liquid nutrient and substrate media are converted to product-containing output liquid streams.

It is a further object of the present invention to provide a method and apparatus by which fungal cell populations may be immobilized and fermentations performed in which liquid nutrient and substrate nutrients are converted to product-containing output liquid streams.

It is a further object of the present invention to provide a method and apparatus by which yeast cell populations may be immobilized and fermentations performed in which liquid nutrient and substrate nutrients are converted to product-containing output liquid streams.

It is a further object of the present invention to provide a method and apparatus by which eukaryotic animal cell populations may be immobilized and fermentations performed in which liquid nutrient and substrate nutrients are converted to product-containing output liquid streams.

It is a further object of the present invention to provide a method and apparatus by which either prokaryotic or eukaryotic plant cell populations may be immobilized and fermentations performed in which liquid nutrient and substrate nutrients are converted to product-containing output liquid streams.

It is a further object of the present invention to provide a method and apparatus by which enzymes or enzyme systems immobilized on solid supports or catalysts immobilized on solid supports or cells or cell components immobilized on solid supports may be immobilized and catalyzed chemical conversions be effected in which liquid substrate nutrients are converted to product-containing output liquid streams.

Another object of the present invention is to provide a method and apparatus by which dissolved oxygen concentrations (or other dissolved gases) in the nutrient liquid flow directed into a cell confinement chamber may be raised to any desired level, depending on the applied hydraulic pressure.

Another object of the present invention is to provide a method and apparatus by which either a nutrient gaseous substrate (such as oxygen) in the nutrient input liquid flow directed into a cell confinement chamber or an excreted respiratory gas (such as, for example, carbon dioxide) in the output liquid flow may be maintained in the dissolved state until liquid-gas disengagement is desired, generally far downstream of the cell immobilization compartment(s).

Another object of the present invention is to provide a method and apparatus by which the conversion of an available chemical substrate into a desired product may be effected by a series of stepwise biocatalyst-mediated conversions in which each chemical conversion step is effected by one of a series of centrifugal immobilization chambers inserted serially or in parallel into the flow stream.

Another object of the present invention is to provide a non-specific, general method and apparatus for cell culture or fermentation which can be applied to any cell type without significant variation.

It is yet another object of the present invention to provide a method and apparatus by which biocatalysts are immobilized within centrifugal immobilization chambers while media containing toxic chemicals are fed into the chambers and the biocatalysts in the chambers neutralize the toxic chemicals thereby converting them into an environmentally benign products.

Another object of the present invention is to provide a method and apparatus for cell culture or fermentation which significantly reduces both the capital and labor costs of production and production facilities.

Another object of the present invention is to provide a method and apparatus for cell culture or fermentation which is much less susceptible to contamination by opportunistic organisms.

Another object of the present invention is to provide a method and apparatus for cell culture or fermentation in which the liquid environment bathing the desired biocatalyst is essentially invariant in time, i.e., the pH, ionic strength, nutrient concentrations, waste concentrations, or temperature do not vary as a function of time in the biocatalyst's environment.

Another object of the present invention is to provide a continuous fermentative or cell culture method.

Another object of the present invention is to provide a method and apparatus for cell culture or fermentation in which cycles of proliferation, growth, or product formation can be accomplished simply by varying the input nutrient feed composition.

Another object of the present invention is to provide a method and apparatus for cell culture or fermentation which can continue for the lifetime(s) of the immobilized micro-organism or cell type.

Another object of the present invention is to provide a method and apparatus for culturing biocatalysts under conditions which thereby significantly increases the yield of products from the biocatalyst.

Another object of the present invention is to provide a method and apparatus for cell culture or fermentation which increases the conversion efficiency (of substrate to product) of the culture process.

Another object of the present invention is to provide a method and apparatus for cell culture or fermentation which significantly reduces the cost of heating or cooling the aqueous media required to support the culture process.

Another object of the present invention is to provide a method and apparatus for cell culture or fermentation which results in higher yields of products such as antibiotics from micro-organism fermentations.

Another object of the present invention is to provide a method and apparatus for cell culture or fermentation which results in higher yields of products such as enzymes or other proteins from micro-organism fermentations.

Another object of the present invention is to provide a method and apparatus for cell culture or fermentation which results in higher yields of products such as ethanol or other short-chain alcohols and acids from the fermentation of micro-organisms.

Another object of the present invention is to provide a method and apparatus for cell culture or fermentation which results in higher yields of products such as protein hormones from genetically-transformed micro-organisms.

Another object of the present invention is to provide a method and apparatus for cell culture or fermentation which results in higher yields of products such as protein hormones from eukaryotic cells.

Another object of the present invention is to provide a method and apparatus for cell culture or fermentation which results in higher yields of products such as amino acids, nitrogenous bases, or alkaloids from the fermentation of micro-organisms.

Another object of the present invention is to provide a method and apparatus for cell culture or fermentation which results in higher yields of products such as fuel-grade ethanol from the fermentation by yeasts of sugar-containing agricultural material.

Another object of the present invention is to provide a method and apparatus which would reduce the fermentation time required to produce alcoholic beverages such as beer and wine.

Another object of the present invention is to provide an easily scaled-up method and apparatus for cell culture or fermentation which can be commercially employed.

These and other objects, features and advantages of the present invention will become apparent after a review of the following detailed description of the disclosed embodiment and the appended claims.

The accompanying figures, which are incorporated and form a part of the specification, illustrate several scientific principles and embodiments of the present invention and, together with the description, serve to explain the principles of the invention.

Fig. 1 illustrates the central features of Counter-Flow Centrifugation.

Fig. 2 illustrates an analysis of the operative forces in Counter-Flow Centrifugation.

Fig. 3 illustrates the central problem with Counter-Flow Centrifugation.

Fig. 4 illustrates the mathematical defect in the conventional treatment of Counter-Flow Centrifugation.

Fig. 5 is an illustration of the effect on immobilized particles using conventional Counter-Flow Centrifugation at long time periods.

Fig. 6 illustrates the modification of Counter-Flow Centrifugation employed in the process of this invention.

Fig. 7 is an illustration of the mathematics governing the motion of a particle due to the effect of gravity on that particle when it is restrained in a centrifugal field exactly opposed by a liquid flow.

Fig. 8 is an illustration of the resultant motion of a particle under the constraints of Fig. 7.

Fig. 9 is a mathematical evaluation of the immobilization conditions at a given radius.

Fig. 10 is an analysis of the balance of centrifugal forces and flow velocity forces in a rotating cylindrical immobilization chamber.

Fig. 11 is an analysis of the balance of centrifugal forces and flow velocity forces in a rotating conical immobilization chamber.

Fig. 12 is an illustration of a three-dimensional array of particles in a rotating conical immobilization chamber.

Fig. 13 is an illustration of the inter-stratum "buffer regions" in a three-dimensional array of particles in a rotating conical immobilization chamber.

Fig. 14 is a mathematical analysis of the intra-stratum flow velocity variation in a two-dimensional array of particles in a rotating conical immobilization chamber.

Fig. 15 is an illustration of an example conical immobilization chamber and the boundary conditions which determine those dimensions.

Fig. 16 is an analysis of the positional variation of the centrifugal and flow velocity forces in the chamber of Fig. 15. at a flow rate of 10 mL/min.

Fig. 17 is a block diagram of a process configuration designed to maintain desired dissolved gas concentrations in the liquid input to a centrifugal bioreactor.

Fig. 18 is an illustration of a liquid flow pressure regulator.

Fig. 19 is a sectional view of an example embodiment of the Centrifugal Fermentation Process when viewed parallel to the axis of rotation.

Fig. 20 is a view of the rotor body of Fig. 19 when viewed parallel to the axis of rotation.

Fig. 21. is a cross-sectional view of one of the demountable biocatalyst immobilization chambers of Fig. 19.

Fig. 22 is a sectional view of the rotor body of Fig. 19. when viewed perpendicular to the axis of rotation.

Fig. 23 is a sectional view of the rotor body of Fig. 19 along the dotted line indicated in Fig. 22, when viewed parallel to the axis of rotation.

Fig. 24 is a sectional view of the rotor body of Fig. 19 along the dotted line indicated in Fig. 22, when viewed parallel to the axis of rotation.

Fig. 25 is a sectional view of the rotor body of Fig. 19 along the dotted line indicated in Fig. 22, when viewed parallel to the axis of rotation.

Fig. 26 is a sectional view of the rotor body of Fig. 19 along the dotted line indicated in Fig. 22, when viewed parallel to the axis of rotation.

Fig. 27 is a sectional view of the rotor body of Fig. 19 along the dotted line indicated in Fig. 22, when viewed parallel to the axis of rotation.

Fig. 28 is an illustration of the axial channels and their termini in the rotating shaft of Fig. 19.

Fig. 29 is a detail view of the distribution hub of the rotating shaft of Fig. 28.

Fig. 30 is a sectional view of a representative high-performance end face seal.

Fig. 3. is a graphical and mathematical representation of the interior compartment of the bioreactor immobilization chamber of Fig. 21.

Fig. 32 is a graph relating the flow rates and rotor speeds which provide for particle immobilization under the dimensional and boundary condition constraints shown on Fig. 15 and for the rotor body of Fig. 19. for particles of sedimentation rates of 0.001 and 0.01 mm/min at flow rates up to 10 mL/min.

Fig. 33 is a graph relating the flow rates and rotor speeds which provide for particle immobilization under the dimensional and boundary condition constraints shown on Fig. 15 and for the rotor body of Fig. 19 for particles of sedimentation rates of 0.1, 1.0, and 10.0 mm/min at flow rates up to 10 mL/min.

Fig. 34 is a graph relating the flow rates and rotor speeds which provide for particle immobilization under the dimensional and boundary condition constraints shown on Fig. 15 and for the rotor body of Fig. 19 for particles of sedimentation rates of 0.1, 1.0, and 10.0 mm/min at flow rates up to 100 mL/min.

Fig. 35 is a graph displaying the relationship of rotor size and volume capacity of embodiments of the process of this invention.

Fig. 36 is a graph displaying the relationship of rotor size and rotational speed required to maintain a Relative Centrifugal Force of 100 X g in embodiments of the process of this invention.

Fig. 37 is a block diagram of a centrifugal process configuration designed to allow serial processing of a precursor chemical through two centrifugal bioreactors.

The development of this immobilization and culture process has its origin in four distinct areas of knowledge. The function of the overall process depends on the use of information from all four areas for its proper function. These areas are: (1) Stoke's Law and the theory of counterflow centrifugation; (2) the geometrical relationships of flow velocity and centrifugal field strength; (3) Henry's Law of Gases; and, (4) the effect of hydraulic pressure on single and multicellular organisms and their cellular or subcellular components.

The central purpose of the process of this invention is immobilization of three-dimensional arrays of particles (cells, subcellular structures, or aggregated biocatalysts) and to provide them with a liquid environment containing dissolved gases which will maximize their viability and productivity. Such cells may include, but are not limited to, a prokaryotic cell, a bacterium, or a eukaryotic cell, such as algae cells, plant cells, yeast cells, fungal cells, insect cells, reptile cells and mammalian cells. The biocatalyst may be, but is not limited to, a subcellular component, an enzyme complex, and/or an enzyme complex immobilized on a solid support.

The dissolved gases of the present invention include but are not limited to air, O₂, NH₃, NO₂, Ar, He, N₂ and H2 or any mixture thereof.

This process utilizes a novel modified form of "Counterflow Centrifugation" to immobilize particle arrays. A proper application of Stoke's Law in combination with provision for the effect of gravity which also acts on the immobilized particles results in a mathematical relationship which allows for the relative immobilization of high-density arrays of such particles. The effect of gravity discussed previously and graphically depicted in Figs. 3-5 can be eliminated by an alternative choice of rotational axis as is shown in Fig. 6. If rotation about the horizontal axis (y) is chosen instead of rotation about the vertical axis (z), as is most common in biological centrifugations, then the effect of gravity on immobilized particles will always be limited to action solely in the x-z plane. Since this is the same plane in which both the centrifugal as well as the liquid flow related forces are constrained to act, the motion of a restrained particle at any point in a rotational cycle is the resultant of the sum of the three types of forces acting upon it.

As is shown in Inset A of Figure 7, where the plane of the Figure is the x-z plane, the effect of gravity (F_g) on the position of a particle suspended in a radially-directed centrifugal field (F_c) while an exactly equal and opposing force supplied by an inwardly-directed flowing liquid (F_b) is directed toward the particle, can be calculated by the evaluation of equations 1-4 where (k) represents the downward displacement in the x-z plane imparted by gravitational forces during an angular rotation of the rotor position equal to (a). Analysis of the motion of a particle under these constraints and for [2π X (k/a)] < R (a low mass particle) results in the determination that the motion is periodic; that is, the particle motion results in a return to its starting place after a complete rotation of 360 degrees (after equilibrium is reached). As is shown in Figure 7, the effect of gravity on the motion of a particle otherwise immobile as a result of the opposing equality of the centrifugal and flow-related forces results in a decrease in radial position in quadrants I and II, and an exactly equal radial lengthening in quadrants III and IV. Thus, the radial distance of the particle from the axis of rotation also exhibits a periodic motion over the course of a full rotation of 360 degrees. It should be noted that, mathematically, measurement of the periodicity of motion requires only one rotation if measurement begins at either 90 or 180 degrees whereas two full rotations are required if measurement begins at either zero or 180 degrees, since a new equilibrium radial distance different from the original results in the latter case.

The effective motion of a particle through a complete rotational cycle is shown in the inset of Figure 8. If the sides of a container in which the particle is suspended are labeled 1 and 2 (see circled numbers in Fig. 8), then the motion of the particle over the course of one rotational cycle would describe a circle with its center displaced toward the "leading edge" side of the particle's container. Thus, a particle suspended in a centrifugal field which is opposed by an equal liquid flow field will be constrained to periodic motion (and thus is effectively immobilized) if the balance of the radially-directed forces can be maintained over the course of its movement.

With these theoretical considerations in mind, we can now return to the hypotheses of Sanderson and Bird which were graphically shown in Figure 2. A corrected graphical representation is shown in Figure 9, in which the axis of rotation is now the (y) axis. Under these conditions the hypothesis of Sanderson and Bird can now be restated and applied to long-term immobilization of particles. Equation 3 of Fig. 9 is now valid. There is a radial distance out the z axis (rz) whose evaluation by Eqn. 3 represents a position where the particle is relatively immobilized in a centrifugal field which is exactly opposed by an inwardly-directed liquid flow, even in the presence of a gravitational field. Furthermore, a simplification of Stoke's Law (Eqn. 1) under the conditions of uniform particle size, shape, and density and a homogeneous liquid flow results in Eqn. 2, where it is obvious that the Sedimentation Velocity of a particle (SV) is a simple linear function of the applied centrifugal field. Similarly, Eqn. 3 can then be rewritten under the same conditions to yield Eqn. 4, where liquid Velocity (V in Eqn. 3) has been replaced by liquid Flow Velocity (FV). Equation 4 suggests that there is a continuum of liquid flow velocities and applied centrifugal fields which could be matched by the evaluation of constant (C), all of which would satisfy the requirement of relative particle immobilization. Further, if the liquid flow velocity could be varied as a function of (z), there could be a separate application of this equation at each radial distance. Consideration of the implications of Eqn. 4 is important for the relative immobilization of three-dimensional arrays of particles as opposed to the immobilization of two-dimensional arrays of particles at a single radial distance from the rotational axis.

If the confinement chamber in which a particle is located is cylindrical (as is graphically depicted in Figure 10) and if a liquid is flowed into this chamber from the end of the chamber most distal to the axis of rotation, then it is obvious that the flow velocity of this liquid flow (as defined in Eqn. 1, Fig. 10) will have a single value at all points not occupied by layers of particles. As a consequence, if a two-dimensional array of particles is in positional equilibrium at a particular radial distance (A₁), as is indicated in Eqn. 2, (where CF is the centrifugal field strength and FV is the liquid flow velocity) then particles forced to occupy positions at radial distances either greater than or smaller than A₁, such as those located in Fig. 10 at A₂ or A₃, will necessarily be presented with an inequality of restraining forces which will result in net translation of the particles. Thus, those particles located at A₂, a longer radial distance than A₁, will experience a greater centrifugal force than those at A₁ and will necessarily migrate to longer radial distances (Eqn. 3). Conversely, particles initially located at A₃ would experience a reduced centrifugal field and would migrate to shorter radial distances (Eqn. 4). Thus, it is not possible to form a three-dimensional array of particles in a "parallel-walled" confinement chamber such as that of Figure 10.

If, however, the confinement chamber has a geometry such that its cross-sectional area increases as the rotational radius decreases, as is graphically displayed in Figure 11, then it is mathematically possible to form three-dimensional arrays of immobilized particles. This is a consequence of the fact that the microscopic flow velocity of the liquid flow varies inversely as the cross-sectional area (Eqn. 1) while the relative centrifugal field varies directly as the rotational radius (Eqn. 2). Thus, if values of flow velocity and rotation velocity are chosen such that a two-dimensional array of particles is immobilized at rotational radius A₁ (Eqn. 3), then it is mathematically possible to adjust the "aspect ratio" of the side walls of the confinement chamber such that those particles initially located at radial distance A₂ could also experience either an similar equality of forces or, as is shown in Eqn. 4, an inequality of forces which results in net motion back toward the center of the chamber. A similar argument may be applied to particles located at A₃ (see Eqn. 5). Although the geometry of the confinement chamber as depicted in Figure 11 is that of a truncated cone, note that other geometries could be alternatively used - subject to the constraint that the cross-sectional area of the chamber increases as the rotational radius decreases. Thus, as is depicted in Figure 12, it is possible to construct a three-dimensional array of particles in a varying centrifugal field opposed by a liquid flow field if the confinement chamber geometry chosen allows for a flow velocity decrease greater than or equal to the centrifugal field strength decrease as the rotational radius decreases. In the geometry chosen in Fig. 12, that of a truncated cone, the two-dimensional arrays of particles at each rotational radius (R_c) will each be constrained to motion toward that radius where the opposing forces are exactly equal.

While, at first glance, the description presented above would suggest that the net effect of the mismatch of forces at all radii other than that which provides immobilization would result in a "cramming" of all particles into a narrow zone centered on the appropriate radius, such is not the case. As is shown graphically in Figure 13, as each layer of particles approaches an adjacent layer, it will move into a region where a "cushioning effect" will keep each layer apart (the horizontal arrows in Fig. 13). The explanation for the inability of adjacent layers of particles to interdigitate is a consequence of an analysis of the microscopic flow velocity profile through each layer. In Figure 14, a single representative stratum of spherical particles confined to a particular radial distance in a chamber layer of circular cross-section is presented. The ratio of the diameters of the particles to the diameter of the cross-section of Fig. 14 is 12:1. While the magnitude of the flow velocity of the liquid through unoccupied portions of the chamber cross-section can be quantified simply from the chamber dimensions at that point, the flow velocity through a region occupied by a stratum of particles will necessarily be much greater than that in their absence as a consequence of the greatly reduced cross-sectional area through which the liquid must travel. As is shown in the graph in Fig. 14, the increase in flow velocity through a stratum of the above dimensions is more than double that determined in the free space just adjacent to the stratum on each side. This microscopic increase in local flow velocity in the region of each stratum effectively provides a "cushion" which keeps each adjacent stratum separate.

In actual use, it has been determined that, for the case of a chamber geometry of a truncated cone, it is preferable that the most distal region of the truncated cone be the region where an exact equality of centrifugal forces and liquid flow velocity be achieved. The "aspect ratio" (the ratio of the lower diameter to the upper diameter) of the conic section is determined by the simultaneous solution of the two equations presented in Figure 15. In Eqn. 2, the desired boundary condition of immobility for that "lowest" stratum of particles is presented. It states that the intrinsic sedimentation rate of the particle due to gravity (S.R.) times the relative centrifugal field applied at that radial distance (RCF) be exactly equal to the magnitude of the liquid flow velocity (F.V.) at that point. In Eqn. 1, a desired boundary condition at the opposite surface of the array of particles is presented. In order to insure retention of all particles within the confinement chamber, a boundary condition wherein the product of S.R. and RCF is twice the magnitude of the flow velocity at that radial distance has been arbitrarily chosen. Simultaneous solution of the desired boundary condition equations is used to solve for the ratio of the conic section diameters when the upper diameter and conic length is given.

Figure 16 is a profile of the relative magnitudes of the flow-related forces and the centrifugal forces across a confinement chamber of conical cross-section which has dimensions in this example of: large diameter = 6.0 cm, small diameter = 3.67 cm, and depth = 3.0 cm. We define the Relative Sedimentation Rate as the product of the intrinsic sedimentation rate of a particle due to gravity in a nutrient media at its optimal temperature and the applied centrifugal field. For a given flow rate (in this example 10 mL/min) into a chamber of the indicated dimensions where the proximal end of the confinement chamber is 9.0 cm from the rotational axis, the product of the intrinsic particle sedimentation rate due to gravity and the angular velocity is a constant at the given flow rate in order to satisfy the desired boundary conditions (see Fig. 15). In other words, the angular velocity needs not to be specified here since its value depends only on the particular particle type it is desired be immobilized. The dotted line displays the linear variation in the centrifugal field strength from the bottom to the top of the confinement chamber, while the solid line displays the corresponding value of the flow velocity. At the bottom of the chamber (its most distal portion), the forces are equal and a particle at this position would experience no net force. At the top of the chamber, a particle would experience a flow-related force which is only one-half of the magnitude of the centrifugal field and would thus be unlikely to exit the chamber, even in the presence of a nearby region of decreasing cross-sectional area (the chamber liquid exit port), where flow velocities will increase markedly.

It should be clear from the foregoing that, subject to the necessary condition that the cross-sectional area increases as rotational radius decreases, there are other geometrical chamber configurations whose shape could be manipulated in order to establish boundary and intermediate relationships between the applied centrifugal field and the liquid flow velocity forces at any radial distance in order to establish desired resultant force relationships in the three-dimensional particle arrays. In practice, however, it is undesirable to utilize geometries with rectangular cross-sections as a result of the anomalous effects of coriolis forces which act in a plane transverse to the rotational plane. In the case of rectangular cross-sections, these otherwise unimportant forces can contribute to interlayer particle motion.

It should also be clear from the foregoing that the effect of gravitational forces acting on the individual particle masses which acts independently of the applied centrifugal forces (see Figs. 7-8) are even less important than was indicated earlier. In particular, since the basic effect of gravity on an otherwise immobilized particle is to either cause radial lengthening or radial shortening, such a motion of a particle will necessarily bring it either into a region of increased flow velocity magnitude (longer radii) or decreased flow velocity magnitude (shorter radii) with only a much smaller change in centrifugal field strength (see Fig. 16). As a consequence, the periodic motion of a particle due to gravitational effects on its intrinsic mass will be severely dampened in the presence of such unbalanced opposing force fields and will amount to, in the case of low mass particles such as biocatalysts, a "vibration in place".

It should also be obvious from the foregoing that there could be, in a practical sense, a severe problem with the maintenance of the immobilized particle arrays in the above fashion when these particles are aerobic cells, micro-organisms, or biocatalytic substructures. Such structures require, in addition to liquid nutrients, the provision of certain nutrients which are gases at ambient temperatures and pressures. For example, the large majority of cells or micro-organisms which are valuable in the production of commercial biochemicals are aerobes, that is, they require oxygen for viability. While these living organisms (or their subcellular constituents) can only utilize oxygen in a dissolved form, the only method of providing it heretofore was by bubbling or sparging the gas through the nutrient liquid in which the cells are suspended in order to effect its solubilization. Further, most living organisms (including certain anaerobes) produce metabolic wastes which are gases (for example, carbon dioxide or methane). If gas volumes were either introduced into or arose from metabolic processes occurring in the immobilized three-dimensional arrays of particles discussed above, then the careful balance of forces which provides for their immobilization would be destroyed.

Thus, the proper function of the centrifugal immobilization process of this invention requires that there be provision made to eliminate the possibility of either the introduction of or the generation of gas(es) within the confinement chamber. Since the only form of these otherwise gaseous chemicals which is utilizable by these cells (or is produced by them) is the aqueous dissolved form, it is this form which must be preserved in the process of this invention. One may insure this condition by the application of Henry's Law, which, in essence, states that the quantity of a gas which may be dissolved in a liquid is a function of the system pressure. Thus, if the hydraulic pressure of the liquid-containing system (the confinement chamber and the liquid lines leading to and from the chamber) are maintained at a hydraulic pressure sufficient to fully dissolve the necessary quantity of input gas and to insure the solubility of any produced gases, then there will be no disturbance of the immobilization dynamics.

Figure 17 is a block diagram which demonstrates one method by which the maintenance of such a gas-free, completely liquid system at hydraulic pressures greater than ambient may be effected. In this system the indicated pumps are all positive displacement pumps; that is, liquid is constrained to motion through the pumps in the directions indicated by the arrows. Pump 3 is the primary feed pump which moves liquids into and out of the cell immobilization chamber which is located in a centrifuge rotor. The raising of the hydraulic pressure in the circuit containing Pump 3 and the cell immobilization chamber is accomplished by placing a liquid pressure regulator, the System Pressure Regulator, at a position in the circuit downstream of the cell immobilization chamber. Thus, the setting of a pressure limit higher than ambient on the system pressure regulator results in no liquid flow through this circuit until the positive displacement pump, Pump 3, moves enough liquid into the circuit to raise the system hydraulic pressure to a value near this setting. Once an equilibrium system pressure is established, the pressurized liquid downstream of Pump 3 will flow continuously at a rate set by control of Pump 3.

In order to dissolve an appropriate amount of a desired nutrient gas into the liquid input to Pump 3, a Gas-Liquid Adsorption Reservoir is placed in the input line leading to Pump 3. Non-gassed liquids are moved from the Media Reservoir into the Gas-Liquid Adsorption Reservoir by means of Pump 1. Quantities of the desired gas (air or oxygen, for example) are, at the same time, let into this reservoir through a pressure regulator set for the gas pressure required to insure the solubilization of the desired concentration of this gas into the nutrient liquid. Note that, in the steady-state, it is necessary that Pump 1 is operated at the same flow rate set for Pump 3. Pump 2 is a recirculation pump which is operated at a flow rate higher than that of Pumps 1 and 3. Pump 2 is used to increase the contact between the gas and liquid phases of the Gas-Liquid Adsorption Reservoir in order that a desired concentration of gas dissolved in the nutrient liquid is maintained in the bulk of the volume of liquid in the adsorption reservoir. It is essential, because of the nature of positive displacement pumps, that the magnitude of the system pressure set with the System Pressure Regulator be at a value higher than the pressure magnitude set in the Gas-Liquid Adsorption Reservoir. In order that a sufficient volume of liquid equilibrated with the desired concentration of gas(es) be available at any time, a valve on the input to Pump 3 may be utilized to allow such equilibration to occur prior to any actual use. Similarly, by means of switching valves, the liquid input to Pump 3 may be changed from that indicated in Figure 17 to any other input reservoir desired, subject to the constraint that the hydraulic pressure of such a reservoir is lower than the value of hydraulic pressure set by the System Pressure Regulator.

Figure 18 is a depiction of a representative, commercially-available liquid pressure regulator. A flow of liquid into the regulator 14 is obstructed by a spring-loaded needle valve 10 which presses against a seat 11. When the hydraulic pressure of the input liquid becomes great enough, the needle valve is displaced from the seat and a flow can then exit the pressure regulator 15. The fixed pressure exerted by the needle valve spring 12 can be adjusted by increasing or decreasing the pressure exerted by the adjustable spring 13.

It should be obvious that the block diagram of Figure 17 is a representation of one of many process flow configurations which may be employed in order to flow a gas-free pressurized liquid through a centrifugal immobilization chamber. In particular, one may envision many different methods of insuring adequate mixing of gas and liquid in order to effect the solubilization of a measured quantity of gas into the liquid. What is central to the process of this invention is: (1) the liquid circuit comprising the centrifugal immobilization chamber and the liquid lines into and out of the chamber be operated at a hydraulic pressure greater than ambient pressure; (2) that there is provision for the solubilization of a desired quantity of a gas into the liquid prior to its insertion into the liquid circuit leading to the centrifugal immobilization chamber(s); and, (3) that the system hydraulic pressure be maintained at a high enough value to keep both the input gas(es), as well as the respiratory gas(es) which may be produced by biological systems in solution throughout the liquid circuit, upstream of the system pressure regulator and downstream of Pump 3. In our hands, hydraulic pressures of 100-2000 psig have proved sufficient to maintain a gas-free liquid environment for all possible conditions of cell density and cell number.

There will be no measurable deleterious effects on the culture of animal cells or micro-organisms or their subcellular constituents as a result of the necessity to increase the hydraulic pressure of their environment in the centrifugal confinement chamber at hydraulic pressures below 10,000 psig. The successful culture of living cells using bioreactor headspace pressurization is a proven and established culture method, albeit limited in scope to pressures of less than 50 psig (see Yang, J. and Wang, N.S. (1992) Biotechnol. Prog. 8, 244-251 and references therein). At hydraulic pressures of 15,000 to 30,000 psig some disassociation of noncovalent protein complexes has been observed, although pressures of more than 90,000 psig are required to denature monomeric proteins (Yarmush, et al. (1992) Biotechnol. Prog. 8, 168-178). It is a seldom appreciated, but well known fact that living cells (and their constituent parts) are unaffected by, and indeed cannot sense hydraulic pressure magnitudes below those limits outlined above. This may best be appreciated in considering the effects of hydraulic pressure on marine organisms. For every 10 meters of depth under the sea, approximately one atmosphere (14.7 psig) of overpressure is gained. Thus, for example, benthic organisms exhibiting biochemical processes and metabolic pathways identical to their shallow-water and terrestrial counterparts inhabit ecological niches and proliferate mightily at hydraulic pressures of more than 3000 pounds per square inch, completely oblivious of this situation. Similarly, the hydraulic pressure under which terrestrial mammalian cells exist is greater than ambient, ranging from ca. 90 to 120 mm Hg greater than ambient in man, for example. The explanation for the "invisibility" of hydraulic pressure in biological systems can be understood if it is realized that hydraulic pressure in aqueous systems has as its "force carrier", the water molecule. Since the lipid bilayer which forms the boundary membrane of living cells is completely permeable to water molecules, an applied hydraulic pressure in aqueous systems is transmitted across the boundary membranes of cells or subcellular organelles by the movement of water molecules with the result that the interior(s) of cells rapidly equilibrate to an externally-applied aqueous hydraulic pressure.

There are situations where hydraulic pressures are deleterious to living cells such as, if a pressure field in an aqueous system is varied at high frequency, then it is possible to cause cell disruption by means of pressure differentials across the cell boundary membrane. However, the frequency required for such lethal effects is quite high, on the order of thousands of cycles per second. As long as the pulsatile pressure of pumping in the process of this invention is kept below such a limit there is no effect on cell viability for even the most fragile of cells as a result of pressure fluctuations. In addition, cell replication is completely unaffected by culture at increased hydraulic pressure.

The problem of the introduction and withdrawal of pressurized liquid flows into and out of a rotating system has been solved by innovations in seal design over the past twenty years. High performance mechanical end-face seals are available which are capable of operation at rotational rates in excess of 5000 revolutions per minute while maintaining a product stream hydraulic pressure of more than 2000 psig, such as those available from Durametallic Corporation (2104 Factory Street, Kalamazoo, MI 49001). Such high performance mechanical seals have leakage rates below 5 liters per year, can be cooled by pressurized refrigerated liquids whose inadvertent leakage into the product stream at the above leakage rates will have no effect on biological systems, and can be operated in a manner which provides for the maintenance of absolute sterility in the product stream. The somewhat inexplicable aversion to the use of mechanical end-face seals for use in centrifugal bioreactor systems (see U.S. Patent Nos. 4,939,087 and 5,151,368, for example) results in a perceived necessity for the connection of flexible tubing (and complicated mechanisms for its "untwisting") in conventional designs. Such designs are, as a result, limited to: (1) hydraulic pressures near one atmosphere as a consequence of tube flexibility requirements; and (2) low rotational speeds and short bioreactor run times as a result of the vigorous motion of these flexing connections. The use of modern high performance mechanical end-face seals eliminate all of these drawbacks to centrifugal bioreactor performance.

Immobilization of three-dimensional arrays of particles in a force field comprised of outwardly directed centrifugal forces which are opposed by inwardly directed liquid flow forces has been described. The effect of gravitational forces which act, inevitably, on even the smallest and lightest of particles over prolonged time periods can be essentially negated and reduced to a small periodic "vibration in place" by the proper choice of rotational axis. The disruptive effects of the possible introduction of gases into this system has been accounted for by raising the hydraulic pressure of the liquid system to values which assure that such otherwise gaseous chemicals will remain dissolved in the flowing liquid. It has been emphasized that the necessary increase in hydraulic pressure will have no effect on biological units such as cells, microorganisms, or their subcellular constituents.

In the following paragraphs, we present a design and an analysis of the performance of that design of the Centrifugal Fermentation Process. Figure 19 depicts the components of an apparatus of this invention. A cylindrical rotor body 20 is mounted on a horizontal, motor-driven rotating shaft 21 inside a safety containment chamber bounded by metal walls 22. The rotor body is fixed in position on the rotating shaft 21 by means of locking collars 23. The rotating shaft is supported on either side of the rotor by bearings 24. The rotating shaft extends outside the safety containment chamber 22 for a distance and ends in a terminal bearing and end cap 29 mounted in an external housing 25. Liquid flows are introduced into and removed from bioreactor chambers 26 mounted in the rotor by means of a liquid input mechanical end-face seal 28 and a liquid output mechanical end-face seal 27 which communicate with liquid channels within the rotating shaft 21. Typical dimensions for an example rotor body (a = 36 cm and b = 15 cm) are entirely reasonable and comparable to rotor dimensions known to those skilled in the art.

Figure 20 is a view of a typical rotor body 20 as viewed parallel to the axis of rotation. The cylindrical rotor body is machined with a shaft mounting channel 30 through its center to allow its mounting on the rotating shaft, is machined to have chamber positioning recesses 32 into which cylindrical demountable bioreactor chambers (26 in Figure 19) may be placed and further has radial rectilinear channels 33 in which metallic liquid lines which communicate with the bioreactors may be located. In actual use, a circular cover (not shown) would be attached to the surface of the rotor shown in Fig. 20 to close the rotor.

Figure 21 is a depiction of one of the demountable bioreactor chambers (shown in Figure 19, 26). The chamber is cylindrical and is composed of two pieces of thick walled metal, a top piece 40, containing a machined conical recess and a machined passage terminating in an output compression fitting 41 by which liquid may be removed from the bioreactor and a lower portion 42 of the same metal which has been internally machined to form the biocatalyst immobilization chamber 43 of a desired geometric shape. The shape of the immobilization chamber depicted in Fig. 21 is that of a truncated cone with a short cylindrical volume at its top face and a short conical volume at its bottom face. A machined passage terminating in an input compression fitting 44 allows liquid input into the immobilization chamber. The two portions of the chamber are bolted together by means of countersunk assembly screws 45 and sealed against an internal positive hydraulic pressure by means of one or more O-ring compression seals 46. In the case of certain animal cell cultures where contact between the immobilized cells and the interior metal walls of the chamber should be avoided, it may be expedient to provide suitable conical inserts of, for example, polyethylene, in order to prevent such contact. Alternatively, the interior of the immobilization chamber might be coated with an appropriate lining material to provide the same effect.

Figure 22 is a transverse sectional view through the rotor body (shown in Figure 19, 20) parallel to the axis of rotation. As is shown in this figure, the bioreactor chambers 26 are connected to centrally located axial channel 50 and eccentric axial liquid input channel 51 within the rotating shaft 21 by means of output liquid transport lines 53 and input liquid transport lines 54. Each liquid transport line is a metal tube which is connected to compression fittings which screw both into machined passages in the rotor shaft and into compression fittings on the bioreactor chambers. As is also shown in Fig. 22, the exact machining of the rotor body may be examined by five different sectional views of the rotor body perpendicular to the axis of rotation (Fig. 23-27 views) which are sectional views at the levels indicated by the dotted lines.

In Figures 23-27 the dimensions and configuration of five different internally machined sections of the rotor body 20 are displayed. Figures 23 and 27 show one method by which the rotor body 20 may be mounted on the rotating shaft, Figure 19, 21, by means of sprocket-shaped recesses 60 concentric with the shaft mounting channel 30 which accept the locating collars (23 in Fig. 19). S-1 is a cross-sectional view of the shaft mounting channel 30 and the sprocket shaped recesses 60. Figure 24 depicts four radial rectilinear channels 33 machined into the rotor body into which the liquid input lines (54 in Fig. 22) will travel. Figure 25 depicts the shapes of the chamber positioning recesses 32 machined into the rotor body into which the cylindrical bioreactor chambers will be placed, and also shows the relationship of these recesses to the radial rectilinear channels 33. Note that the radial rectilinear channels 33 extend farther radially than do the chamber positioning recesses 32 and thus provide a support channel against which the liquid lines will rest as they extend "upward" to connect with an input compression fitting (44 Figure 21) of the bioreactor chambers (see Figure 22). By virtue of the fact that the input liquid line (a length of metal tubing) is supported by its resting against a wall of the most distal radial rectilinear channel 33 as it makes the right angle bend to travel to its terminus at an input compression fitting 44 of each bioreactor chamber (see section S-2, Fig. 24), there will be no extra centrifugal stress applied to these lines as a result of the rotational movement of the system.

Figure 26 details the internal machining of the rotor body 20 for liquid output line attachment recesses 70 necessary in order to provide working room for the mechanical attachment of the the output liquid transport lines 53 of Figure 22, to an output compression fitting 41, Figure 21, of the bioreactor chambers. As is shown in Fig. 22, the output liquid transport lines 53 are bent into a "U-shaped" configuration (exaggerated in the figure) which allows their length to be adjusted during mechanical connection to the output compression fittings of the bioreactor chambers. The bioreactor chambers 26 are supported against centrifugal stress by the distal walls of the chamber positioning recesses 32; no weight is imparted to the output liquid lines (except their own) as a result of centrifugal forces.

Figure 28 is a view of that portion of the rotating shaft 21 on which the rotor body 20 is mounted and that portion of the rotating shaft on which the liquid output mechanical end-face seal 27 and the liquid input mechanical end-face seal 28 which convey liquid flows into and out of the bioreactor chambers 26 are mounted. The rotating shaft 21 contains two liquid transport axial channels, the centrally located axial channel 51, and the eccentric axial liquid input channel 50. The centrally located axial channel 51 transports the liquid output of the bioreactor chambers to the liquid output mechanical end-face seal 27 by means of short radially-directed connecting passage 82 while the eccentric axial liquid input channel 50 conveys liquid from the liquid input mechanical end-face seal 28 to the bioreactor chamber, also by means of a short radially-directed connecting passage 81. The centrally located axial channel 51 and eccentric axial liquid input channel 50 extend from one end of the shaft to the region where the rotor body will be located. Compression plugs 80 seal the terminal axial openings of these channels.

Figure 29 is a view of the radially-disposed liquid distribution channel hubs in the region of the rotating shaft 21 where the rotor body will be mounted. Two pairs of channels, the radial output liquid line channels 90 and the radial input liquid line channels 92 are machined through two cross-sections of the shaft. These channels are, in the case of the radial output liquid line channels 90, directly in communication with the centrally located axial channel 51. In the case of the radial input liquid line channels 92, an additional radial passage 94 is machined which connects the eccentric axial liquid input channel 50 with the central connection of the radial input liquid line channels 92. This additional radial passage 94 is sealed with a compression plug 95 at the surface of the rotating shaft 21. In actual practice, particularly in the case of high speed operation of the rotor system, it may be preferable that both axial passages, the centrally located axial channel 51 and the eccentric axial liquid input channel 50, be eccentric to the axis of rotation and located symmetrically on a diameter of the rotating shaft for balancing purposes.

Figure 30 is a view of a mechanical end-face seal assembly, such as the liquid output mechanical end-face seal 27, shown in Figure 19. The liquid output mechanical end-face seal 27 is mounted on the rotating shaft 21 and positioned with an opening to the interior liquid space of the seal over a short radially-directed connecting passage 82 which communicates with the centrally located axial channel 51 machined into the rotating shaft 21. A seal between the rotating and stationary portions of the liquid output mechanical end-face seal 27 is provided by the contact of the stationary seal face 100 against the rotating seal face 102. In the case of high performance end-face seals utilizable in the process of this invention, where consideration must be made for the resultant centrifugal forces which act on the seal components, all spring elements are located in the stationary portion of the seal assembly. While the seal configuration shown in Fig. 30 is that of a single seal, double and/or tandem end-face seal configurations may prove more advantageous in prolonged usage. Not shown in the figure are pressurized cooling liquid passages and jacketing necessary to maintain temperature equilibrium in the seal assembly. When such a seal assembly is mounted on the rotating shaft of the present invention, aqueous liquids may be pumped into the stationary part of the seal assembly via compression fitting attachment and the pumped liquid will follow the path indicated by the dotted line 103 to make communication with the centrally located axial channel 51 which transports this liquid away from the bioreactor chambers 26 mounted in the rotor body 20 (see Figure 28).

The principal disadvantage heretofore in the employment of mechanical seals for the transfer of liquids into and out of rotating systems, where the purpose of the system is to culture biological entities such as animal cells or micro-organisms, has been the problem of the maintenance of sterility. Low pressure mechanical seals have, in the past, provided a route by which adventitious micro-organisms can gain entrance into bioreactor systems via the thin film of internal liquid which lubricates the end-face seal surfaces. In the process of this invention, where the internal liquid is always held at a hydraulic pressure higher than ambient, all leakage of liquid will occur to the exterior of the system; there is thus no possible route through which adventitious contaminants can enter the system. Furthermore, the small leakage of internal nutrient or product liquid which might exit the bioreactor system through the mechanical seals of the process of this invention (which might, for example, contain micro-organisms in certain applications) will not be free to dissipate into the environment. As a consequence of the operating characteristics of high-speed, high-pressure mechanical seals, it will be necessary to surround the seal components with a pressurized cooling liquid flow. It has been found, in practice, that an ideal liquid which possesses the proper viscosity and flow properties for the cooling of such seals is 75-85% glycerol. Any leakage of internal liquid to the exterior in the process of this invention will result in its dispersion into the body of this recirculated liquid. We have found that glycerol at this concentration is completely unable to support the growth of a number of representative animal cells or micro-organisms; this is likely a general phenomenon, presumably as a result of the osmotic movement of water out of the living cells into the glycerol. Thus, periodic sanitary disposal of the cooling liquid volume of glycerol when it becomes diluted with leakage volumes and its replacement with fresh glycerol will serve to maintain sterility in the single place in the system where liquids might escape. Finally, since it is possible, after prolonged use, that loss of internal system pressure or incipient failure of the seal systems might allow liquid flow in the reverse direction across the seal faces, it is important to note that small quantities of glycerol which could thus leak into the bioreactor system would not be anything but an additional nutrient when diluted into the flowing internal process liquid.

In order to obtain data for an analysis of the performance of a rotor body of the dimensions indicated by the letters placed on Figs. 19, 20, 22-29 and containing demountable cylindrical bioreactor immobilization chambers like that depicted in Fig. 21, (dimensions denoted by letters), it was necessary that several scale dimensions and boundary equations be chosen arbitrarily and used to determine operating characteristics of a typical example embodiment of the present invention. The immobilization boundary equations are those listed in Equations 1 and 2 of Figure 15. The rotor dimensions chosen for this example and indicated by letter in Figures 19-29 are as follows: a 15.0 cm j 10.0 cm s 2.54 cm b 36.0 cm k 1.50 cm t 4.0 cm c 1.27 cm l 6.0 cm u 6.14 cm d 1.0 cm m 0.5 cm v 1.0 cm e 1.73 cm n 1.0 cm w 1.0 cm f 3.0 cm o 1.0 cm x 6.5 cm g 7.0 cm p 5.0 cm y 5.0 cm h 2.0 cm q 6.0 cm z 5.5 cm i 2.0 cm r 4.0 cm

In this example, the geometry of the interior immobilization region 43 of the example bioreactor chamber 26 is that of a truncated cone. As is shown in Figure 31, the dimensional problem of the "aspect ration" of the interior immobilization region of the example bioreactor chamber due to the boundary condition constraints can be reduced to an examination of the geometrical relationships between the large and small radii of the conical section and the height of the conical section.

As is shown in Fig. 31A, the conical immobilization region of an example bioreactor chamber (when viewed in the plane of rotation) can be depicted as a truncated triangle 110 whose proximal face is at a distance of R_x from the center of rotation and whose truncated length is R_c. Centrifugal forces (RCF) act to cause translation of a particle 111 in the immobilization volume to longer radii, while liquid flow forces (FV) act to cause translation to shorter radii. Equation (1) of Fig. 31B is an expression for the magnitude of the Relative Centrifugal Force at radial length (Rx) in terms of the rate of rotation. Equation (2) is an expression for the magnitude of the flow velocity at radial length (R_x) in terms of the liquid flow rate and the large diameter of the conic section. Equation (3) is an expression for the magnitude of the Relative Centrifugal Force at radial length (R_x + R_c) in terms of the rate of rotation. Equation (4) is an expression for the magnitude of the flow velocity at this latter radial length in terms of the liquid flow rate and the given dimensions of the depicted triangle and its sections. In order to determine the "aspect ratio" of the conic section (the ratio of the diameters) which will satisfy certain boundary conditions, given the physical dimensions of the rotor body, we have chosen to express the radius of the narrow end of the conic section in terms of the length (L) of the cone from which this section is derived.

The desired boundary conditions are: (I) that the product of the intrinsic sedimentation rate of the immobilized particle due to gravity (SR) and the applied centrifugal field (RCF) be exactly equal to the magnitude of the flow-derived force (FV) at the most distal portion of the chamber; and, (2) that this product be twice the magnitude of the liquid flow forces at the most proximal portion of the chamber. Thus:$${\text{at centrifugal radius = R}}_{\text{x}}{\text{+ R}}_{\text{c}}$$$$\text{(SR) X (RCF) = FV}$$$${\text{at centrifugal radius = R}}_{\text{x}}$$$$\text{(SR) X (RCF) = 2 X FV}$$

Substituting into these equations the dimensional specifications for RCF and FV obtained from Eqns. (1 - 4) of Fig. 31, we now have two simultaneous equations which relate the liquid flow rate, the speed of rotation of the system, and the dimensions of the conical immobilization chamber:$${\text{(1)\hspace{1em}\hspace{1em}\hspace{1em}(SR) C}}_{\text{1}}{\text{(R}}_{\text{x}}{\text{+ R}}_{\text{c}}{\text{) = C}}_{\text{2}}\text{(}\frac{\text{L}}{{\text{L - R}}_{\text{C}}}{\text{)}}^{\text{2}}$$$${\text{(2)\hspace{1em}\hspace{1em}\hspace{1em}(SR) C}}_{\text{1}}{\text{(R}}_{\text{x}}{\text{) = 2 X C}}_{\text{2}}$$

In order to arrive at a solution to these equations, we will make the following substitutions which are based on the physical dimensional limits of the example rotor system:$${\text{R}}_{\text{x}}\text{= 90 mm}$$$${\text{R}}_{\text{c}}\text{= 30 mm}$$$$\text{q = 30 mm}$$

The simultaneous equations now become:$${\text{(1)\hspace{1em}\hspace{1em}\hspace{1em}(SR) C}}_{\text{1}}{\text{(120) = C}}_{\text{2}}\text{(}\frac{\text{L}}{\text{L - 30}}{\text{)}}^{\text{2}}$$$${\text{(2)\hspace{1em}\hspace{1em}\hspace{1em}(SR) C}}_{\text{1}}{\text{(90) = 2 X C}}_{\text{2}}$$

Substituting Eqn. (2) into Eqn. (1) yields:$$\text{(}\frac{\text{L}}{\text{L - 30}}{\text{)}}^{\text{2}}\text{= 240/90}$$

Solution of this quadratic expression yields L = 77.4 mm and:$$\text{(}\frac{\text{L}}{\text{L - 30}}{\text{)}}^{\text{2}}\text{= 2.67}$$

Since it was earlier determined (see Fig. 31) that:$${\text{R}}_{\text{1}}\text{=}\frac{{\text{q (L - R}}_{\text{c}}\text{)}}{\text{L}}$$

Thus, the smaller radius of the truncated cone which satisfies the boundary conditions is:$${\text{R}}_{\text{1}}\text{= 18.4 mm}$$

Now, the two simultaneous equations become:$${\text{(1)\hspace{1em}\hspace{1em}\hspace{1em}(SR) C}}_{\text{1}}{\text{(120) = C}}_{\text{2}}\text{(2.67)}$$$${\text{(2)\hspace{1em}\hspace{1em}\hspace{1em}(SR) C}}_{\text{1}}{\text{(90) = C}}_{\text{2}}\text{(2)}$$    and, by subtracting (2) from (1) and collecting terms, we arrive at:$${\text{(3)\hspace{1em}\hspace{1em}\hspace{1em}(SR)(30)C}}_{\text{1}}{\text{= (0.67)C}}_{\text{2}}$$

Substitution into (3) of the values of C₁ and C₂ yields:

Now we have an expression which satisfies the desired boundary conditions and physical dimensional constraints in terms of the controllable variables, RPM and FR:$$\sqrt{\text{SR}}\text{(RPM) = (2.65)}\sqrt{\text{FR}}$$

Thus, once the physical dimensions of the rotor system as well as those of the bioreactor immobilization chamber have been determined, the range of rotor angular velocities (RPM) and the system liquid flow rates (FR) which will constrain the particles to immobility in the bioreactor will follow a simple relationship which is dependent only on the intrinsic sedimentation rate (SR) of the object particle due to gravity. Note that, under the above conditions, the maximal volume of immobilization is ca. 56 mL per bioreactor chamber.

This method and apparatus for containing a biocatalyst comprises the step of containing the biocatalyst in at least one bioreactor chamber placed in a centrifugal force field wherein a continuous flow of a liquid acts to create a force which opposes the centrifugal force field and wherein a gravitational force contributes to the resultant vector summation of all forces acting on the biocatalyst, wherein the gravitational force, the centrifugal force and the opposing liquid substantially immobilize the biocatalyst at a position in the chamber, such that the effect of the gravitational force on the biocatalyst is essentially negated and reduced to a periodic vibration, the liquid at hydraulic pressures greater than the ambient barometric pressure, and wherein no gaseous phase is present in said chamber

There are many alternative shapes for the bioreactor chambers which are contemplated in this invention. Such an alternative embodiment is a bioreactor chamber having its inner space in the shape of a right circular cone whose major axis is aligned in parallel with the applied centrifugal force field and whose large diameter is nearer to the axis of rotation than is its apex.

Another alternative embodiment is the bioreactor chamber having its inner space in the shape of a right circular cone whose major axis is not aligned in parallel with the applied centrifugal force field but presents an angle between 0 and 90 degrees to the applied centrifugal force field. Also included in the present invention is the bioreactor chamber inner space having its inner space in the shape of a truncated right circular cone whose major axis is aligned in parallel with the applied centrifugal force field and whose large diameter is nearer to the axis of rotation than is its minor diameter.

Additionally, the present invention includes the bioreactor chamber having its inner space in the shape of a truncated right circular cone whose major axis is not aligned in parallel with the applied centrifugal force field but presents an angle between 0 and 90 degrees to the applied centrifugal force field.

The present invention includes the bioreactor chamber having its inner space in the shape of a sphere where the applied centrifugal force field is perpendicular to a circular cross-section of the spherical bioreactor chamber and a chamber where the applied centrifugal force field is not perpendicular to a circular cross-section of the spherical bioreactor' chamber but presents an angle between 0 and 90 degrees to the circular cross-section.

Additionally, the present invention includes a bioreactor chamber having its inner space is in the shape of a truncated sphere where the applied centrifugal field is perpendicular to a circular cross-section of the sphere and a chamber where the applied centrifugal field is not perpendicular to a circular cross-section of the sphere but presents an angle between 0 and 90 degrees to the circular cross-section.

The present invention includes a bioreactor chamber having its inner space in a shape which possesses a varying circular cross-section where the applied centrifugal force field is perpendicular to the circular cross-sections and a chamber where the applied centrifugal force field is not perpendicular to the circular cross-sections but presents an angle between 0 and 90 degrees to the circular cross-sections.

Additionally, present invention includes a bioreactor chamber having its inner space in a shape which possesses a varying elliptical cross-section where the applied centrifugal force field is perpendicular to the elliptical cross-sections and a chamber where the applied centrifugal force field is not perpendicular to the elliptical cross-sections but presents an angle between 0 and 90 degrees to the elliptical cross-sections.

Also included in present invention is the bioreactor chamber having its inner space in a shape which is a combination of circular and elliptical cross-sections along an axis perpendicular to the applied centrifugal force field and a chamber has its inner space in a shape which is a combination of circular and elliptical cross-sections along an axis which is not perpendicular to the applied centrifugal forcefield but presents an angle between 0 and 90 degrees to the circular and/or elliptical cross-sections.

The process of this invention is directed toward the immobilization of biocatalysts such as micro-organisms, eukaryotic cells, their subcellular organelles, and natural or artificial aggregates of such biocatalysts. Thus, the process system must be capable of immobilizing fairly light particles. It is known that the sedimentation rates of such particles due to gravity range from ca. 0.01 mm/min for small bacteria to 0.1 mm/min for small animal cells to more than 10.0 mm/min for thick-walled micro-organisms (such as yeasts) and biocatalytic aggregates such as bead-immobilized cells. We have analyzed the performance characteristics of the centrifugal bioreactor system of this invention using the dimensional configurations outlined above and present these results below.

Figure 32 displays profiles of the values of rotor speed and liquid flow rates which satisfy the boundary conditions outlined earlier and for the rotor and bioreactor dimensions outlined in Figs. 19-27 in the cases of typical biologically significant particles of the lowest two sedimentation rate ranges. The upper line displays the continuum of liquid flow rates and rotor speeds which result in the immobilization of particles of an intrinsic sedimentation rate of 0.001 mm/min, a value smaller by a factor of ten than any we have measured for any tested micro-organism. Note that, even at a flow rate of 10 mL/min, the rotor speed required to maintain immobilization is a physically reasonable value and that the maximum centrifugal force required is ca. 9400 X g, a value well within the physical limits of average quality centrifugal systems. The lower line displays the corresponding profile for particles of a sedimentation rate of 0.01 mm/min, a value near that exhibited by typical representative bacteria. Again, this line represents a continuum of values which satisfy the immobilization conditions. Thus for example, if a flow rate of 2.0 mL/min is required to adequately nutrition a particular sized three-dimensional array of "bacteria A", a rotor speed near 1200 rpm will suffice, while a required flow rate of 8.0 mL/min necessitates a rotor speed near 2500 rpm. Note that the heavier particles of SR = 0.01 mm/min require only a modest maximal centrifugal force of ca. 1000 X g at a flow rate of 10.0 mL/min.

Figure 33 displays profiles of the values of rotor speed and liquid flow rates which satisfy the boundary conditions outlined earlier and for the rotor and bioreactor dimensions outlined in Figs. 19-27 in the cases of typical biologically significant particles of the higher three sedimentation rate ranges. The upper line displays the continuum of liquid flow rates and rotor speeds which result in the immobilization of particles comparable to larger micro-organisms or small animal cells (for example, mammalian erythrocytes) of an intrinsic sedimentation rate of 0.1 mm/min. The middle line displays the corresponding values for the immobilization of more typical animal cells (ca. 30 µm diameter), while the bottom line displays the continuum of values which provide for the immobilization of large dense cells, such as eukaryotic yeasts. As was the trend shown in Fig. 32, it is obvious from the data of Fig. 33 that the maximum rotor speeds and maximal centrifugal forces required in this flow rate range get smaller as the intrinsic particle sedimentation rate due to gravity increases. Thus, for a flow rate of 10.0 mL/min, a three-dimensional array of average-sized animal cells requires only a relative centrifugal force of ca. 10 X g to provide immobilization.

Figure 34 displays profiles of the values of rotor speed and liquid flow rates which satisfy the boundary conditions outlined earlier and for the rotor and bioreactor dimensions outlined in Figs. 19-27 in the case of liquid flow rates of as much as 100 mL/min for the highest three intrinsic sedimentation rate ranges examined. Thus, even if the liquid flow rates required to nutrition such immobilized "beds" of particles (example bed volume = 56 mL) is increased ten-fold, the maximal centrifugal forces and rotor speeds required are technically unremarkable. Note that, in the case of "animal cells" (SR = 1.0 mm/min) a flow rate of 100 mL/min represents a flow of 6.0 L/hr, a flow obviously larger than that required to adequately nutrition such a three-dimensional array of cells under any imaginable conditions.

While it is generally obvious that the effect of immobilizing a population of, for example, bacteria in a flowing liquid will not lead to cellular damage as a result of the flow of liquid past the surface of such cells (since many micro-organisms possess extracellular "sheaths" which protect their plasma membranes from liquid shear forces), it is less obvious whether or not delicate animal cells (which do not possess such extracellular protection) would remain viable under these conditions. However, as was shown in Fig. 33, the maximum RCF required to maintain an average-sized animal cell immobile in a liquid flow of 10 mL/min is ca. 10 X g. Even if this flow is raised to a level decidedly well above any anticipated nutritional need (100 mL/min), the maximun RCF required is only ca. 100 X g (Fig. 34). It should be remembered that the immobilization of such a cell in a flowing liquid is the mathematical equivalent of moving the cell through a stationary liquid. Thus, since the conventional laboratory sedimentation of animal cells through liquid media at RCF's of more than 100 X g is an unremarkable phenomenon, it is unlikely that the shear forces acting on such cells in the process of this invention will cause any damage to their plasma membranes. This assertion is supported by the operating characteristics of a related device, the Beckman JE-5.0 Centrifugal Elutriation System, from which viable animal cells have been successfully recovered after exposure to flow rates and RCFs greatly in excess of those proposed herein for the process of this invention.

With the present invention, it is possible to immobilize three-dimensional arrays of biologically-significant particles and to adequately nutrition the immobilized particles with a completely liquid flow. In particular, for the small-scale prototypic centrifugal process outlined above, the required centrifugal forces and liquid flow rates are not unusual and present no novel problems such as, for example, requiring unreasonably high rotational speeds or flow rates. Further, it has been demonstrated that there is a wide range of paired flow rate and angular velocity values which maintain the immobilization of three-dimensional arrays of such particles.

The fact that there is a wide range of flow rates and corresponding rotational speeds which can be used to immobilize such arrays of particles has, however, a wider significance. It is safe to say that, using conventional culture methodology, the major problem encountered in large-scale culture is the inability to adequately nutrition dense masses of metabolically-active biological units. In the case of conventional mammalian cell culture for example, an average cell density of more than 1 X 10⁶ cells/mL is rarely achieved for prolonged time periods for this reason. Similarly, bacterial cell densities of 1 X 10⁷ cells/mL are rarely exceeded in mass culture by conventional methods for this same reason. Using the methodology of the process of this invention, as cell density and effective "bed" volume increases (either from cellular proliferation or bioreactor loading), the increased nutritional requirements of larger or more dense cultures can be met by increasing input liquid flow while simultaneously increasing the size of the applied centrifugal field. Using the process of this invention, it is possible to easily maintain mammalian cell cultures at concentrations two powers of ten greater than conventional, with bacterial cell densities approaching 1 X 10¹⁰ cells/mL equally realizable.

Similarly, for dense cultures of aerobic organisms, the conventional problem of adequate delivery of optimal dissolved oxygen to the culture is easily solved using the process of this invention. Since it is possible to dissolve molecular oxygen in typical culture media at concentrations of more than 100 mM (using a hydraulic pressure of ca. 1500 psig) the problem of the delivery of optimal dissolved oxygen, for any imaginable dense culture, is solved simply by adjusting the system hydraulic pressure to a value which will maintain the solubility of the desired concentration of oxygen. The ability to maintain dissolved oxygen concentration at optimal levels results in greatly increased production efficiency. As has been noted by many researchers, the inability to achieve cellular production efficiencies near those observed in vivo is a major disadvantage of conventional animal culture techniques (The Scientist, 8, #22, pg.16, November 14, 1994). Furthermore, the ability to achieve near-normal aerobic efficiency in dense culture has another less obvious advantage: the generation of heat. Instead of requiring expensive energy input to bring the liquid cellular environment to an optimal temperature, it is likely that the pumped liquid of the process of this invention will have to be delivered to the cellular environment at reduced temperatures in order to carry away excess metabolic heat.

Another important advantage of the process of this invention is the relative invariance of the chemical composition of the liquid environment in which the three-dimensional arrays of biocatalysts are immobilized. Since the arrays are continually presented with fresh, optimal liquid nutrient input and since these arrays are continually drained by the continuance of the process flow, the chemical composition of the cellular environment will be completely invariant in time. There will be shallow chemical gradients of nutrients, product(s), and metabolites across the radial length of these arrays, but since this is the shortest dimension of the array, these gradients will be minimal and can be easily compensated for by tailoring of the media composition. Thus, for example, a pH change across the array depth can be compensated for with minimal buffering while input nutrient gradients across the array depth can be similarly compensated for. Most important, however, is the fact that metabolic waste products will be continually removed from the cellular environments by the liquid process flow. Since it has been suggested that the inability to remove metabolic wastes and the inability to continually remove desired products from the cellular environment is a major factor in lowered per-cell productivity, it is likely that the utilization of the process of this invention will markedly increase general cellular productivity.

The chemical composition of optimal input liquid nutrient media to immobilized populations of biocatalysts in the process of this invention will be quite different from that of conventional nutrient media. In particular, the optimal media composition in this process will be that which can be completely consumed in one pass through the immobilization bioreactor. Typical nutrient media contains a mix of as many as thirty or more nutrient chemicals, all of which are in great excess over the immediate nutritional needs of the biocatalysts, since this media must sustain their metabolic processes for as long as 100 hours in some cases. Similarly, conventional media contains concentrations of pH buffer compounds and indicators and hormonal stimuli (fetal sera and/or cytokines, etc.) in great excess over the immediate needs of the biocatalysts. In the process of this invention, the input liquid medium can be tailored to contain those concentrations of nutrients and stimulants which are directly required by the immediate metabolism of the immobilized biocatalysts. Ideally, the outflowing liquid which exits the bioreactor would be completely devoid of nutrients and contain only salts, metabolic wastes, and product molecules. The possibility of tailoring the input media in order to maintain an immobilized cellular population in a nutritional state which either promotes or inhibits cellular proliferation would thus be realizable. It is highly unlikely that a nutritional mix which is optimal for cellular division is optimal for the production of biochemicals by cells at rest in the cell cycle.

The liquid medium may be any formulation known to those skilled in the art or may include specific individual components which are necessary for the biocatalyst of interest. The kinds of media may include, but are not limited to, a nutrient medium, a balanced salt solution, or a medium containing one or more organic solvents. The medium may contain dissolved gases for growth of the biocatalyst under anaerobic or aerobic conditions. The medium may be formulated so that the biocatalyst product or mobile biocatalysts found in the medium are more easily isolated.

Another less obvious implication of the utility of this process methodology is the effect of scaling. In the prototypic example outlined earlier, the total volume capacity of the four-bioreactor rotor is ca. 224 mL. Note however, that as the radius of the rotor is increased the volume capacity of the system goes up as the cube of the radius. This is graphically shown in Figure 35 where the lowest point on this graph corresponds to the prototypic example (total volume = 225 mL). A rotor with a radius of 1.5 meters would have a volume capacity of ca. 120 liters. Further, since the average density of culture is roughly 100 times that of conventional culture methods, the equivalent culture volume is proportionally larger. Thus, a centrifugal fermentation unit with a rotor radius of 1.5 m is roughly equivalent to a 12,000 liter fermentation using current technology. Finally, it should be noted that there is an additional advantage in scale in the use of the process of this invention. As a consequence of the fact that relative centrifugal force is directly proportional to the radial length but is also directly proportional to the square of the angular velocity, the rotational speeds required decrease as the rotational radius is increased for a given centrifugal field strength. This is shown graphically in Figure 36. While the rotational speed required to maintain a RCF = 100 X g is ca. 810 rpm for a rotor with a radius of 18 cm, this required rotational speed drops to less than 300 rpm when the rotational radius is increased to 1.5 m, more than a 50% lowering in the speed of rotation.

While it is obvious that scale-up of this process will have value in industrial production facilities, it should be noted that a miniature embodiment of the Centrifugal Fermentation Process could be valuable in the analytical study of the "metabolic physiology" of small homogeneous populations of a particular cell type. To our knowledge, the exact nutritional requirements for maximal proliferation of, for example, a bacterial population are unknown - and could be rapidly and easily determined by perturbation of the composition of the nutritional liquid input to an immobilized test population while measuring some output parameter indicative of growth. Similarly, while it is desirable to know exactly what nutritional mix is optimal for cellular production of a biological product (a nutritional mix which is highly unlikely to be identical to that which maximizes proliferation), such parameters are, again, unknown. We believe that small-scale versions of the process of this invention could be advantageously utilized in advancing "analytical microbiology" or "analytical cell biology" in a fashion heretofore impossible to perform.

The most immediately obvious usage for the method of this invention is likely for the continuous production of biological products which are secreted or otherwise released into the out-flowing liquid stream. Thus, for example, one might utilize this process for the continual harvest of product(s) released from an immobilized micro-organism population whose growth rate (and death rate) have been nutritionally manipulated to maintain a steady state immobilized "bed volume". Such a process could run, theoretically, forever. Similarly, the immobilization of secretory animal cell populations would result in continual outflow of liquid enriched in the desired product(s).

In the case of non-secreted products (such as the cytosolic accumulation of protein in genetically-engineered E. coli), the process of this invention will also be extremely useful. If an immobilized cell population is maintained in the bioreactor system outlined above, but under conditions of excess nutritional input, then the population will quickly grow to an enlarged bed size which will continually "leak" excess cells into the out-flowing liquid stream. Thus, the process of this invention can be operated as a "production cow", that is, as a continual incubator for the production and outflow of mature cells rich in the desired product. Downstream isolation and disruption of the out-flowing cell stream to capture the product of interest would then follow conventional product purification methods.

The process of this invention offers the possibility of continual, serial interconversion of bio-organic substrates through several intermediate steps by two or more separate animal cell populations or micro-organism populations. As a consequence of the ability of the process of this invention to completely immobilize biocatalyst populations while continually flowing a liquid stream into and out of the immobilized population, it now becomes possible to serially connect separate, disparate immobilized populations into one flowing process stream with the assurance that there will be no cross-contamination of one population with the other. To acomplish this, several of the devices described herein are connected in series so that materials flow from one device into another device and then into the following device and so on.

As is shown in Figure 37, a process flow schematic in which a biochemical substrate provided as a dissolved nutrient in the primary media reservoir is converted into intermediate "product A" by its passage through the biocatalyst population immobilized in Centrifuge and Rotor #1 and is then further converted into "product B" by passage through a biocatalyst population immobilized in Centrifuge and Rotor #2. Furthermore, it is possible to change the composition of the liquid nutritional feedstock between the two immobilized populations since neither centrifugal immobilization chamber is constrained to operate at the same flow rate and angular velocity as the other. Thus, as is shown in Fig. 37, the liquid flow into Centrifuge and Rotor #2 may be modified by means of an additional pump supplying necessary nutrients from Media Reservoir #2; the total flow per unit time through Centrifuge and Rotor #2 is simply higher than that through Centrifuge and Rotor #1.

A commercially-valuable example of the utility of such a serial conversion process of this type is the biological production of acetic acid. Anaerobic bio-conversion of glucose into ethanol by an immobilized population of a yeast such as Saccharomyces cerevisiae in Centrifuge and Rotor #1 could be followed by aerobic conversion of ethanol to acetic acid by an immobilized population of the bacterium Acetobacter acetii located in Centrifuge and Rotor #2, if dissolved oxygen and supplemental nutrients are provided via Media Reservoir #2 (using, for example, the oxygenation scheme depicted in Fig. 17).

In a similarly fashion, if a process flow scheme demanded that total flow volume per unit time through specific centrifugal bioreactor units be reduced, then a series of identical centrifugal bioreactor units could be connected in parallel to the process stream flow, with the resultant individual flow volume per unit time thereby reduced to the fractional flow through each unit. In this case, the devices of the present invention would be connected in a parallel arrangement.

The microbial organisms which may be used in the present invention include, but are not limited to, dried cells or wet cells harvested from broth by centrifugation or filtration. These microbial cells are classified into the following groups: bacteria, actinomycetes, fungi, yeast, and algae. Bacteria of the first group, belonging to Class Shizomycetes taxonomically, are Genera Pseudomonas, Acetobacter, Gluconobacter, Bacillus, Corynebacterium, Lactobacillus, Leuconostoc, Streptococcus, Clostridium, Brevibacterium, Arthrobacter, or Erwinia, etc. (see R.E. Buchran and N.E. Gibbons, Bergey's Manual of Determinative Bacteriology, 8th ed., (1974), Williams and Wilkins Co.). Actinomycetes of the second group, belonging to Class Shizomycetes taxonomically, are Genera Streptomyces, Nocardia, or Mycobacterium, etc. (see R.E. Buchran and N.E. Gibbons, Bergey's Manual of Determinative Bacteriology, 8th ed., (1974), Williams and Wilkins Co.). Fungi of the third group, belonging to Classes Phycomycetes, Ascomycetes, Fungi imperfecti, and Bacidiomycetes taxonomically, are Genera Mucor, Rhizopus, Aspergillus, Penicillium, Monascus, or Neurosporium, etc. (see J.A. von Ark, "The Genera of Fungi Sporulating in Pure Culture", in Illustrated Genera of Imperfect Fungi, 3rd ed., V. von J. Cramer, H.L. Barnett, and B.B. Hunter, eds. (1970), Burgess Co.). Yeasts of the fourth group, belonging to Class Ascomycetes taxonomically, are Genera Saccharomyces, Zygosaccharomyces, Pichia, Hansenula, Candida, Torulopsis, Rhodotorula, Kloechera, etc. (see J. Lodder, The Yeasts: A Taxonomic Study, 2nd ed., (1970), North-Holland). Algae of the fifth group include green algae belonging to Genera Chlorella and Scedesmus and blue-green algae belonging to Genus Spirulina (see H. Tamiya, Studies on Microalgae and Photosynthetic Bacteria, (1963) Univ. Tokyo Press). It is to be understood that the foregoing listing of micro-organisms is meant to be merely representative of the types of micro-organisms that can be used in the fermentation process according to the present invention.

The culture process of the present invention is also adaptable to plant or animal cells which can be grown either in monolayers or in suspension culture. The cell types include, but are not limited to, primary and secondary cell cultures, and diploid or heteroploid cell lines. Other cells which can be employed for the purpose of virus propagation and harvest are also suitable. Cells such as hybridomas, neoplastic cells, and transformed and untransformed cell lines are also suitable. Primary cultures taken from embryonic, adult, or tumorous tissues, as well as cells of established cell lines can be employed. Examples of typical such cells include, but are not limited to, primary rhesus monkey kidney cells (MK-2), baby hamster kidney cells (BHK21), pig kidney cells (IBRS2), embryonic rabbit kidney cells, mouse embryo fibroblasts, mouse renal adenocarcinoma cells (RAG), mouse medullary tumor cells (MPC-11), mouse-mouse hybridoma cells (I-15 2F9), human diploid fibroblast cells (FS-4 or AG 1523), human liver adenocarcinoma cells (SK-HEP-1), normal human lymphocytic cells, normal human lung embryo fibroblasts (HEL 299), WI 38 or WI 26 human embryonic lung fibroblasts, HEP No. 2 human epidermoid carcinoma cells, HeLa cervical carcinoma cells, primary and secondary chick fibroblasts, and various cell types transformed with, for example, SV-40 or polyoma viruses (WI 38 VA 13, WI 26 VA 4, TCMK-1, SV3T3, etc.). Other suitable established cell lines employable in the method of the present invention will be apparent to the person of ordinary skill in the art.

The products that can be obtained by practicing the present invention are any metabolic product that is the result of the culturing of a cell, either eukaryotic or prokaryotic; a cell subcellular organelle or component, such as mitochondria, nuclei, lysozomes, endoplasmic reticulum, golgi bodies, peroxisomes, or plasma membranes or combinations thereof; or an enzyme complex, either a natural complex or a synthetic complex, i.e., a plurality of enzymes complexed together to obtain a desired product.

One of the advantages of the present invention is the ability to produce a desired chemical from a cell without having to go through the laborious process of isolating the gene for the chemical and then inserting the gene into a suitable host cell so that the cell (and thus the chemical) can be produced in commercial quantities. By practicing the present invention, a mammalian cell that is known to produce a desired chemical can be directly cultured in high density according to the present invention to produce large quantities of the desired chemical.

Products that can be produced according the present invention include, but are not limited to, immunomodulators, such as interferons, interleukins, growth factors, such as erythropoietin; monoclonal antibodies; antibiotics from micro-organisms; coagulation proteins, such as Factor VIII; fibrinolytic proteins, such as tissue plasminogen activator and plasminogen activator inhibitors; angiogenic proteins; and hormones, such as growth hormone, prolactin, glucagon, and insulin.

The term "culture medium" includes any medium for the optimal growth of microbial, plant, or animal cells or any medium for enzyme reactions including, but not limited to, enzyme substrates, cofactors, buffers, and the like necessary for the optimal reaction of the enzyme or enzyme system of choice. Suitable culture media for cell growth will contain assimilable sources of nitrogen, carbon, and inorganic salts, and may also contain buffers, indicators, or antibiotics.

Any culture medium known to be optimal for the culture of microorganisms, cells, or biocatalysts may be used in the present invention. While such media are generally aqueous in nature for the culture of living organisms, organic solvents or miscible combinations of water and organic solvents, such as dimethylformamide, methanol, diethyl ether and the like, may be employed in those processes for which they are proved efficacious, such as those bioconversions in which immobilized biocatalysts are employed. Passage of the liquid media through the process system may be either one-pass or the liquid flow may be recycled through the system for higher efficiency of conversion of substrate to product. Desired nutrients and stimulatory chemicals may be introduced into the process flow, either via the low pressure nutrient supply or via injection into the process flow upstream of the cell chamber.

It will be appreciated that the present invention is adaptable to any of the well-known tissue culture media including, but not limited to, Basal Medium Eagle's (BME), Eagle's Minimum Essential Medium (MEM), Dulbecco's Modified Eagle's Medium (DMEM), Ventrex Medium, Roswell Park Medium (RPMI 1640), Medium 199, Ham's F-10, Iscove's Modified Dulbecco Medium, phosphate buffered salts medium (PBS), and Earle's or Hank's Balanced Salt Solution (BSS) fortified with various nutrients. These are commercially-available tissue culture media and are described in detail by H.J. Morton (1970) In Vitro 6, 89-108. These conventional culture media contain known essential amino acids, mineral salts, vitamins, and carbohydrates. They are also frequently fortified with hormones such as insulin, and mammalian sera, including, but not limited to, bovine calf serum as well as bacteriostatic and fungistatic antibiotics.

Although cell growth or cell respiration within the centrifugal cell chamber cannot be directly visualized, such metabolism may be readily monitored by the chemical sensing of substrate depletion, dissolved oxygen content, carbon dioxide production, or the like. Thus, for example in the case of a fermentation of a species of Saccharomyces cerevisiae, inoculation of the immobilization chamber with a small starter population of cells can be followed by an aerobic fermentation regime in which glucose depletion, dissolved oxygen depletion, and carbon dioxide production across the cell confinement chamber are measured either chemically or via appropriate sensing electrodes. Thus, cell replication can be allowed to proceed until an optimal cell bed size is reached. Withdrawal of dissolved oxygen input at this time causes the immobilized yeast cells to shift into anaerobic fermentation of glucose with a resultant production of ethanol, a process which can likewise be monitored chemically.

Similarly, without any process modification, the process of the present invention can be utilized as a bioreactor for immobilized chemical catalysts, enzymes or enzyme systems. In such a process, a catalyst, an enzyme or an enzyme system is chemically immobilized on a solid support including, but not limited to, diatomaceous earth, silica, alumina, ceramic beads, charcoal, or polymeric or glass beads which are then introduced into the bioreactor immobilization chamber. The reaction medium, either aqueous, organic, or mixed aqueous and organic solvents, flows through the process system and through the three-dimensional array of solid supports within the bioreactor. The catalyst, enzyme, or enzyme system converts a reactant in the process flow medium into the desired product or products. Similarly, in other applications, either cells or cell components including, but not limited to, vectors, plasmids, or nucleic acid sequences (RNA or DNA) or the like may be immobilized on a solid support matrix and confined for similar utilization in converting an introduced reactant into a desired product.

Commercial application of the present invention can be in the production of medically-relevant, cellularly-derived molecules including, but not limited to, anti-tumor factors, hormones, therapeutic enzymes, viral antigens, antibiotics and interferons. Examples of possible product molecules which might be advantageously prepared using the method of the present invention include, but are not limited to, bovine growth hormone, prolactin, and human growth hormone from pituitary cells, plasminogen activator from kidney cells, hepatitis-A antigen from cultured liver cells, viral vaccines and antibodies from hybridoma cells, insulin, angiogenisis factors, fibronectin, HCG, lymphokines, IgG, etc. Other products will be apparent to a person of ordinary skill in the art.

The foregoing description of the present invention is not intended to be exhaustive or to limit the invention to the precise form disclosed. Many modifications and variations are possible. The particular embodiment described is intended to best explain the principles of the invention and its practical application to thereby enable others skilled in the relevant art to best utilize the present invention in various embodiments and with various modifications as are suited to the particular use contemplated. It is intended to cover in the appended claims all such modifications, variations, and changes as fall within the scope of the process of this invention.